CN112801188A - NILMD method based on affine propagation clustering algorithm and selective Bayesian classification - Google Patents
NILMD method based on affine propagation clustering algorithm and selective Bayesian classification Download PDFInfo
- Publication number
- CN112801188A CN112801188A CN202110127482.7A CN202110127482A CN112801188A CN 112801188 A CN112801188 A CN 112801188A CN 202110127482 A CN202110127482 A CN 202110127482A CN 112801188 A CN112801188 A CN 112801188A
- Authority
- CN
- China
- Prior art keywords
- selective
- sample target
- sample
- bayesian classification
- nilmd
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 title claims abstract description 19
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 33
- 238000009826 distribution Methods 0.000 claims abstract description 10
- 238000007476 Maximum Likelihood Methods 0.000 claims abstract description 9
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000013016 damping Methods 0.000 claims description 3
- 238000012804 iterative process Methods 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 230000010355 oscillation Effects 0.000 claims description 3
- 230000000087 stabilizing effect Effects 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 230000000644 propagated effect Effects 0.000 claims 4
- 238000011161 development Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000007654 immersion Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
- G06F18/24155—Bayesian classification
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an NILMD method based on affine propagation clustering algorithm and selective Bayesian classification, which comprises the following steps: calculating the running state of a sample target based on an AP algorithm; performing probability fitting on the running state of the sample target based on selective Bayesian classification to obtain a running state interval of the sample target so as to realize discretization of the running state of the sample target; and based on a two-dimensional Gaussian distribution and maximum likelihood estimation method, the sample target is identified and decomposed by combining the running state interval of the sample target, and the non-invasive load identification and decomposition of the sample target are completed. The invention solves the problem of poor recognition and decomposition precision of the traditional load decomposition method, and carries out sub-invasive load recognition and decomposition on the sample target by combining the AP algorithm and the selective Bayesian classification method, thereby realizing high-precision recognition and decomposition and improving the working efficiency.
Description
Technical Field
The invention relates to the technical field of electric power, in particular to an NILMD (network independent distance decomposition) method based on an affine propagation clustering algorithm and selective Bayesian classification.
Background
With the improvement of the power utilization level of industries and residents, the power utilization characteristics of the urban power grid demand side show a diversified development trend, and the influence of load differentiation characteristics including power utilization time intervals, load sizes, evaluation indexes, curve forms and the like on power grid scheduling and maintenance is increased year by year.
In the above sea power grid as an example, the construction background of 'digital pump power' and the strategic guidance of 'one, two, six and six' are considered, and in order to promote the urban energy Internet state development and the company digital transformation, the intellectualization and refinement level of power grid dispatching is comprehensively improved, and the non-invasive identification and decomposition for the resident load is the current important requirement.
In the eighties of the last century, George w.hart proposed a non-invasive load monitoring and decomposition (NILMD) method for more conveniently obtaining power consumption data of various electrical appliances of users. Firstly, to obtain the electrical data of a user bus, relevant data monitoring equipment is installed at an inlet of electrical equipment of a user, then the data is analyzed, and the running state of the electrical appliance is analyzed by means of a load identification algorithm.
However, with the development of intelligent measurement technology and the rapid increase of power consumption information on the load side, the existing load decomposition method cannot meet the actual requirement, and a new method is needed to realize the NILMD.
Disclosure of Invention
The invention aims to provide an NILMD method based on an affine propagation clustering algorithm and selective Bayesian classification. The method aims to solve the problem that the traditional load decomposition method is poor in identification and decomposition precision, and the sample target is subjected to sub-invasive load identification and decomposition by combining an AP algorithm and a selective Bayesian classification method, so that high-precision identification and decomposition are realized, and the working efficiency is improved.
In order to achieve the above purpose, the present invention provides an NILMD method based on an affine propagation clustering algorithm and selective bayesian classification, comprising the following steps:
step 1: calculating the running state of a sample target based on an affine propagation clustering (AP) algorithm;
step 2: performing probability fitting on the running state of the sample target based on selective Bayesian classification to obtain a running state interval of the sample target so as to realize discretization of the running state of the sample target;
and step 3: and based on a two-dimensional Gaussian distribution and maximum likelihood estimation method, the sample target is identified and decomposed by combining the running state interval of the sample target, and the non-invasive load identification and decomposition of the sample target are completed.
Most preferably, the AP algorithm comprises the steps of:
step 1.1: constructing a characteristic sample set of the sample target based on the low-frequency characteristics of the sample target;
step 1.2: updating the responsibility and the credibility of the characteristic sample set;
step 1.3: stabilizing oscillation of the characteristic sample set in the iterative process through a damping factor;
step 1.4: and repeating iteration on the characteristic sample set to obtain the running state of the characteristic sample set.
Most preferably, constructing a sample set of characteristics of the sample object comprises the steps of:
step 1.1.1: searching a clustering center of each data point of a sample target in the network and a membership between the data point and the data center through a message transfer mechanism;
step 1.1.2: and dividing the sample targets according to the membership between the data center and the vertex to form a characteristic sample set.
Most preferably, before performing the probability fitting, the method further comprises: and performing probability calculation on the feature sample set and the category variable to obtain the posterior probability of the feature sample set.
Most preferably, the probability fitting by the selective bayesian classification is performed by a bayesian classifier (SBC).
Most preferably, the identification of the sample objects comprises the steps of:
step 3.1: based on two-dimensional Gaussian distribution, performing stochastic description on the operation state interval to construct a power fluctuation model;
step 3.2: and identifying and decomposing the sample target based on a maximum likelihood estimation method and in combination with a power fluctuation model to finish non-invasive load identification and decomposition.
By using the method, the problem of poor recognition and decomposition precision of the traditional load decomposition method is solved, and the sample target is subjected to sub-invasive load recognition and decomposition by combining an AP algorithm and a selective Bayesian classification method, so that high-precision recognition and decomposition are realized, and the working efficiency is improved.
Compared with the prior art, the invention has the following beneficial effects:
the NILMD method provided by the invention solves the problem of poor identification and decomposition precision of the traditional load decomposition method, and carries out sub-invasive load identification and decomposition on the sample target by combining the AP algorithm and the selective Bayesian classification method, thereby realizing high-precision identification and decomposition and improving the working efficiency.
Drawings
FIG. 1 is a schematic flow diagram of the NILMD method provided by the present invention.
Detailed Description
The invention will be further described by the following specific examples in conjunction with the drawings, which are provided for illustration only and are not intended to limit the scope of the invention.
The invention provides an NILMD method based on affine propagation clustering algorithm and selective Bayesian classification, as shown in figure 1, comprising the following steps:
step 1: calculating the running state of a sample target based on an affine propagation clustering (AP) algorithm and combined with continuous variable state electrical appliances with concentrated load characteristics;
wherein the affine propagation clustering (AP) algorithm comprises the following steps:
step 1.1: based on the low-frequency characteristics of the sample target with active power and reactive power as main components, constructing a characteristic sample set X of the sample target, wherein X is { X }1,x2,…,xn};
In this embodiment, a sample set X ═ X of characteristics of the sample object is defined1,x2,,xnThe sequence is the total power sequence P of the sample target with the time period T(j)Extracting power sequences P of M sample targets; wherein the content of the first and second substances,total power sequence P of sample targets of time period T(j)Satisfies the following conditions:
the power sequences P of the M sample targets satisfy:
P=[P1,P2,…,PT];
in the present embodiment, the time period T is 10 days; the sample target is a household appliance; sample targets included three household appliances, kitchen refrigerator, electric fan and immersion heater.
Wherein, constructing a characteristic sample set X ═ { X of the sample object1,x2,…,xnThe method specifically comprises the following steps:
step 1.1.1: searching a clustering center of each data point of a sample target in the network and a membership between the data point and the data center through a message transfer mechanism;
step 1.1.2: dividing sample targets according to membership between the data center and the vertexes to form a plurality of characteristic sample sets X ═ X with specific significance1,x2,…,xn};
Wherein the dividing the sample target comprises: let similarity matrix Kn,nThe median h of (a) is a self-reference and is assigned to Kn,nDiagonal elements of (a).
Step 1.2: for any two samples X in the characteristic sample set XiAnd xjUpdating the responsibility r (i, j) and the credibility a (i, j) respectively; wherein, the update responsibility r (i, j) specifically satisfies:
where s (i, j) is any two samples X in the feature sample set XiAnd xjNegative value of euclidean distance of; a (i, j') is the confidence level after updating; j' is the running state of the characteristic sample set;
the update reliability a (i, j) specifically satisfies:
step 1.3: stabilizing oscillation of the characteristic sample set X in the iterative process by a damping factor lambda to accelerate convergence; the mth iteration of the feature sample set X satisfies:
rm(i,j)=λrm-1(i,j)+(1-λ)rm(i,j)
am(i,j)=λam-1(i,j)+(1-λ)am(i,j)
step 1.4: repeating iteration on the characteristic sample set X to obtain the running state j' of the characteristic sample set X, finishing iteration and realizing state discretization of a sample target; wherein, the running state j' of the characteristic sample set X satisfies the following conditions:
all data points in the AP algorithm are potential clustering centers, the granularity of clustering results can be controlled by adjusting the self-reference degree, the larger the self-reference degree value is, the more the number of clustering results is, and the smaller the self-reference degree value is, the smaller the number of clustering results is.
Step 2: based on selective Bayesian classification, performing probability fitting on the operation state j' of the sample target to obtain an operation state interval to which the operation power of the sample target belongs so as to realize discretization of the operation state of the sample target;
wherein, based on the selective Bayesian classification, the method also comprises the following steps before probability fitting: for the feature sample set X ═ X1,x2,…,xnAnd l class variables CiPerforming probability calculation to obtain posterior probability P (C) of the characteristic sample set Xi| x), and satisfies:
wherein P (X) is X ═ X1,x2,…,xn-probability of co-occurrence; p (C)i) Is a prior probability of class i, which is equal to the ratio of the number of samples for which i belongs to the class to the total number of samples, and satisfies:
where p (x) is a constant, the classification result is not affected and can be ignored, and therefore, the above formula can be expressed as:
for feature xjThe conditional probability P (x) is calculated by the following formulaj|Ci):
In the formula niThe number of samples belonging to the category i in the sample set; k (x)j,xj,i,m,θi) As a nuclear parameter of thetaiA gaussian kernel function of; x is the number ofj,i,mIs xjThe value of the mth iteration in category i.
In the embodiment, the probability fitting of the selective Bayesian classification is realized by a Bayesian classifier (SBC); a Selective Bayesian Classifier (SBC) calculates the conditional probability of a feature by a gaussian kernel based kernel density estimation method, i.e. fitting a probability density function using a superposition of multiple gaussian distributions.
And step 3: and based on a two-dimensional Gaussian distribution and maximum likelihood estimation method, the sample target is identified and decomposed by combining the running state interval of the sample target, and the non-invasive load identification and decomposition of the sample target are completed.
Wherein, identifying the sample target respectively comprises the following steps:
step 3.1: the steady state power of the discretized sample target continuously changes in an operation state interval with the state power as the center, and the continuous change follows Gaussian distribution; the randomness of active power and reactive power under different running states is described based on two-dimensional Gaussian distribution, and a power fluctuation model f (X) is constructedT) (ii) a And the power fluctuation model f (X)T) Satisfies the following conditions:
wherein cov (X) is a covariance matrix; x is an observation vector; mu is an observation vector mean value;
step 3.2: based on maximum likelihood estimation method, combining power fluctuation model f (X)T) And identifying and decomposing the sample target to complete non-invasive load identification and decomposition.
In the present embodiment, the power fluctuation function of the qth household appliance among the Q household appliances at the time t isBased on a maximum likelihood estimation method, the power decomposition satisfies the following steps:
wherein, P(q)The active power of the qth household appliance at the moment t; q(q)The reactive power of the qth household appliance at the moment t; p(q)allThe active power of the total load at the moment; q(q)allIs thatReactive power of the total load at that moment.
The working principle of the invention is as follows:
calculating the running state of a sample target based on an AP algorithm; performing probability fitting on the running state of the sample target based on selective Bayesian classification to obtain a running state interval of the sample target so as to realize discretization of the running state of the sample target; and based on a two-dimensional Gaussian distribution and maximum likelihood estimation method, the sample target is identified and decomposed by combining the running state interval of the sample target, and the non-invasive load identification and decomposition of the sample target are completed.
In conclusion, the NILMD method based on the affine propagation clustering algorithm and the selective Bayesian classification solves the problem of poor recognition and decomposition precision of the traditional load decomposition method, and the sample target is subjected to sub-invasive load recognition and decomposition by combining the AP algorithm and the selective Bayesian classification method, so that high-precision recognition and decomposition are realized, and the working efficiency is improved.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.
Claims (6)
1. An NILMD method based on affine propagation clustering algorithm and selective Bayesian classification is characterized by comprising the following steps:
step 1: calculating the running state of the sample target based on an affine propagation clustering algorithm;
step 2: performing probability fitting on the operation state based on selective Bayesian classification to obtain an operation state interval of a sample target;
and step 3: and based on a two-dimensional Gaussian distribution and a maximum likelihood estimation method, the sample target is identified and decomposed by combining the operation state interval, and the non-invasive load identification and decomposition of the sample target are completed.
2. The NILMD method based on affine propagation clustering algorithm and selective bayesian classification according to claim 1, wherein said affine propagation clustering algorithm comprises the following steps:
step 1.1: constructing a characteristic sample set of the sample target based on the low-frequency characteristics of the sample target;
step 1.2: updating the responsibility and the credibility of the characteristic sample set;
step 1.3: stabilizing oscillation of the feature sample set in the iterative process through a damping factor;
step 1.4: and repeating iteration on the characteristic sample set to obtain the running state of the sample target.
3. The NILMD method based on affine propagated clustering algorithm and selective bayesian classification as claimed in claim 2, wherein constructing the feature sample set comprises the steps of:
step 1.1.1: searching a clustering center of each data point of a sample target in the network and a membership between the data point and the data center through a message transfer mechanism;
step 1.1.2: and dividing the sample targets according to the membership to form the characteristic sample set.
4. The NILMD method based on affine propagated clustering algorithm and selective bayesian classification as claimed in claim 1, wherein prior to performing said probability fitting further comprises: and performing probability calculation on the feature sample set and the category variable to obtain the posterior probability of the feature sample set.
5. The NILMD method based on affine propagated clustering algorithm and selective Bayesian classification as claimed in claim 1, wherein said selective Bayesian classification for said probability fitting is implemented by a Bayesian classifier.
6. The NILMD method based on affine propagated clustering algorithm and selective bayesian classification according to claim 1, wherein said identification of sample objects comprises the following steps, respectively:
step 3.1: based on two-dimensional Gaussian distribution, performing stochastic description on the operation state interval to construct a power fluctuation model;
step 3.2: and identifying and decomposing the sample target by combining the power fluctuation model based on a maximum likelihood estimation method to finish non-invasive load identification and decomposition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110127482.7A CN112801188A (en) | 2021-01-29 | 2021-01-29 | NILMD method based on affine propagation clustering algorithm and selective Bayesian classification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110127482.7A CN112801188A (en) | 2021-01-29 | 2021-01-29 | NILMD method based on affine propagation clustering algorithm and selective Bayesian classification |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112801188A true CN112801188A (en) | 2021-05-14 |
Family
ID=75812950
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110127482.7A Pending CN112801188A (en) | 2021-01-29 | 2021-01-29 | NILMD method based on affine propagation clustering algorithm and selective Bayesian classification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112801188A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7979363B1 (en) * | 2008-03-06 | 2011-07-12 | Thomas Cecil Minter | Priori probability and probability of error estimation for adaptive bayes pattern recognition |
CN109492667A (en) * | 2018-10-08 | 2019-03-19 | 国网天津市电力公司电力科学研究院 | A kind of feature selecting discrimination method for non-intrusive electrical load monitoring |
CN111428816A (en) * | 2020-04-17 | 2020-07-17 | 贵州电网有限责任公司 | Non-invasive load decomposition method |
CN111553444A (en) * | 2020-05-15 | 2020-08-18 | 东南大学 | Load identification method based on non-invasive load terminal data |
CN111564842A (en) * | 2020-06-03 | 2020-08-21 | 吉林大学 | Method for statistical estimation of medium-and-long-term energy consumption in non-invasive electric load monitoring |
-
2021
- 2021-01-29 CN CN202110127482.7A patent/CN112801188A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7979363B1 (en) * | 2008-03-06 | 2011-07-12 | Thomas Cecil Minter | Priori probability and probability of error estimation for adaptive bayes pattern recognition |
CN109492667A (en) * | 2018-10-08 | 2019-03-19 | 国网天津市电力公司电力科学研究院 | A kind of feature selecting discrimination method for non-intrusive electrical load monitoring |
CN111428816A (en) * | 2020-04-17 | 2020-07-17 | 贵州电网有限责任公司 | Non-invasive load decomposition method |
CN111553444A (en) * | 2020-05-15 | 2020-08-18 | 东南大学 | Load identification method based on non-invasive load terminal data |
CN111564842A (en) * | 2020-06-03 | 2020-08-21 | 吉林大学 | Method for statistical estimation of medium-and-long-term energy consumption in non-invasive electric load monitoring |
Non-Patent Citations (4)
Title |
---|
《电工技术学报》: "基于紧邻传播聚类和遗传优化的非侵入式负荷分解方法", 《电工技术学报》, pages 3868 - 3878 * |
庄卫金;张鸿;方国权;陈中;: "基于设备运行状态挖掘的非侵入式负荷分解方法", 电力建设, no. 08 * |
江帆等: "基于选择性贝叶斯分类的非侵入式负荷识别方法", 《电力建设》, pages 94 - 99 * |
闭应洲等 著: "《数据挖掘与机器学习》", vol. 2020, 31 January 2020, 浙江科学技术出版社, pages: 67 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111353653B (en) | Photovoltaic output short-term interval prediction method | |
Cai et al. | Predicting the energy consumption of residential buildings for regional electricity supply-side and demand-side management | |
Kang et al. | A weight-incorporated similarity-based clustering ensemble method based on swarm intelligence | |
Jiang et al. | Inbound tourism demand forecasting framework based on fuzzy time series and advanced optimization algorithm | |
CN106446967A (en) | Novel power system load curve clustering method | |
Ye et al. | Combined approach for short-term wind power forecasting based on wave division and Seq2Seq model using deep learning | |
CN105404783A (en) | Blind source separation method | |
CN109255726A (en) | A kind of ultra-short term wind power prediction method of Hybrid Intelligent Technology | |
CN111563827B (en) | Load decomposition method based on physical characteristics of electrical appliance and electricity utilization behaviors of residents | |
CN109242136A (en) | A kind of micro-capacitance sensor wind power Chaos-Genetic-BP neural network prediction technique | |
Jie et al. | Review on the research of K-means clustering algorithm in big data | |
CN110263834A (en) | A kind of detection method of new energy power quality exceptional value | |
CN113065704A (en) | Hyper-parameter optimization and post-processing method of non-invasive load decomposition model | |
CN114240687A (en) | Energy hosting efficiency analysis method suitable for comprehensive energy system | |
Wang et al. | Sustainable investment forecasting of power grids based on the deep restricted Boltzmann machine optimized by the lion algorithm | |
CN113537556A (en) | Household short-term load prediction method based on state frequency memory network | |
CN109359671B (en) | Classification intelligent extraction method for hydropower station reservoir dispatching rules | |
CN112801188A (en) | NILMD method based on affine propagation clustering algorithm and selective Bayesian classification | |
Wang et al. | Prediction of heating load fluctuation based on fuzzy information granulation and support vector machine | |
Tang et al. | A novel hybrid FA-Based LSSVR learning paradigm for hydropower consumption forecasting | |
CN115186882A (en) | Clustering-based controllable load spatial density prediction method | |
Qin | Software reliability prediction model based on PSO and SVM | |
Chen et al. | Short-term power load model based on combined optimization of cuckoo algorithm and lightGBM | |
Zhang et al. | A novel hybrid forecasting system based on data augmentation and deep learning neural network for short-term wind speed forecasting | |
Cao et al. | A dynamic neural network optimization model for heavy metal content prediction in farmland soil |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |